Abstract
It is unclear whether grammar can influence visual processing of non-linguistic stimuli in a causal manner. If yes, how deep is this effect? Here, an experimental group was trained on a semi-artificial grammar featuring two novel markers of verb transitivity, while a control group was exposed to the same materials but with the novel markers used interchangeably. After that, both groups performed two visual oddball tasks with non-linguistic stimuli that were disguised as a totally irrelevant study. The first (attentional) oddball task showed that the novel grammar cancelled an attentional bias against the dimension highlighted by the grammar (i.e. motion transitivity), as indexed by the P300 amplitude; the second (pre-attentive) oddball task showed that the grammar increased subjects’ pre-attentive response to sudden changes in motion transitivity, as indexed by the N2/visual mismatch negativity (vMMN).